Face Generation¶

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data¶

You'll be using two datasets in this project:

MNIST
CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)

Found mnist Data
Found celeba Data

Explore the Data¶

MNIST¶

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')

<matplotlib.image.AxesImage at 0x7fd9419ea6a0>

CelebA¶

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))

<matplotlib.image.AxesImage at 0x7fd941921c50>

Preprocess the Data¶

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network¶

You'll build the components necessary to build a GANs by implementing the following functions below:

model_inputs
discriminator
generator
model_loss
model_opt
train

Check the Version of TensorFlow and Access to GPU¶

This will check to make sure you have the correct version of TensorFlow and access to a GPU

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))

TensorFlow Version: 1.1.0
Default GPU Device: /gpu:0

Input¶

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
Z input placeholder with rank 2 using z_dim.
Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    # TODO: Implement Function
    inputs_real = tf.placeholder(tf.float32, (None , image_width, image_height, image_channels), name='input_real')
    inputs_z = tf.placeholder(tf.float32, (None, z_dim), name='input_z')
    lr = tf.placeholder(tf.float32, shape=(), name='learning_rate')
    
    return inputs_real, inputs_z, lr


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)

Tests Passed

Discriminator¶

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the discriminator, tensor logits of the discriminator).

def discriminator(images, reuse=False, alpha = 0.2):
    """
    Create the discriminator network
    :param images: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    drop = 0.1
    with tf.variable_scope('discriminator', reuse=reuse):
        #print ('Images:',images.shape)
        #Input is 28x28x3
        x1 = tf.layers.conv2d(images, 64, 5, strides=2, kernel_initializer=tf.contrib.layers.xavier_initializer_conv2d(),
                              padding='same', use_bias=False, activation=None)
        x1 = tf.layers.batch_normalization(x1, training=True)
        x1 = tf.layers.dropout(x1, rate=drop, training=True)
        relu1 = tf.maximum(alpha * x1, x1)
        #relu1 = tf.nn.dropout(relu1, drop)
        #16x16x32
        #print ('L1:',relu1.shape)
        
        x2 = tf.layers.conv2d(relu1, 128, 5, strides=2, kernel_initializer=tf.contrib.layers.xavier_initializer_conv2d(),
                              padding='same',use_bias=False, activation=None)
        bn2 = tf.layers.batch_normalization(x2, training=True)
        bn2 = tf.layers.dropout(bn2, rate=drop, training=True)
        relu2 = tf.maximum(alpha * bn2, bn2)
        #relu2 = tf.nn.dropout(relu2, drop)
        
        #8x8x128
        #print ('L2:',relu2.shape)
        
        x3 = tf.layers.conv2d(relu2, 256, 5, strides=2, kernel_initializer=tf.contrib.layers.xavier_initializer_conv2d(),
                              padding='same',use_bias=False, activation=None)
        bn3 = tf.layers.batch_normalization(x3, training=True)
        bn3 = tf.layers.dropout(bn3, rate=drop, training=True)
        relu3 = tf.maximum(alpha * bn3, bn3)
        #relu3 = tf.nn.dropout(relu3, drop)
        #print ('L3:',relu3.shape)
        
        #Flatten 
        flat = tf.reshape(relu3, (-1, 4*4*256))
        #print ('Out:', flat.shape)
        logits = tf.layers.dense(flat, 1)
        out = tf.sigmoid(logits)
        

    return out, logits


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)

Tests Passed

Generator¶

Implement generator to generate an image using z. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

def generator(z, out_channel_dim, is_train=True, alpha = 0.2):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    drop = 0.5
    xavier = tf.contrib.layers.xavier_initializer_conv2d()
    with tf.variable_scope('generator', reuse=(not is_train)):
        #print ('z:',z.shape)
        x1 = tf.layers.dense(z, 4*4*256)
        x1 = tf.reshape(x1, (-1, 4, 4, 256))
        bn1 = tf.layers.batch_normalization(x1, training=is_train)
        bn1 = tf.layers.dropout(bn1, rate=drop, training=True)
        relu1 = tf.maximum(alpha * bn1, bn1)
        
        #print ('GL1:',relu1.shape)
        #256
        x2 = tf.layers.conv2d_transpose(relu1, 128, 4, strides=1, kernel_initializer=tf.contrib.layers.xavier_initializer_conv2d(),
                              padding='valid')
        #x2 = tf.layers.dropout(x2, rate=drop, training=True)
        bn2 = tf.layers.batch_normalization(x2, training=is_train)
        relu2 = tf.maximum(alpha * bn2, bn2)
        #print ('GL2:',relu2.shape)
        #128
        x3 = tf.layers.conv2d_transpose(relu2, 64, 5, strides=2, kernel_initializer=tf.contrib.layers.xavier_initializer_conv2d(),
                              padding='same')
        bn3 = tf.layers.batch_normalization(x3, training=is_train)
        bn3 = tf.layers.dropout(bn3, rate=drop, training=True)
        relu3 = tf.maximum(alpha * bn3, bn3)
        #print ('GL3:',relu3.shape)
        
        #Output
        logits = tf.layers.conv2d_transpose(relu3, out_channel_dim, 5, strides=2, padding='same') 
        #print ('Out:',logits.shape)
        out = tf.tanh(logits)
    
    return out


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)

Tests Passed

Loss¶

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

discriminator(images, reuse=False)
generator(z, out_channel_dim, is_train=True)

import numpy as np

def model_loss(input_real, input_z, out_channel_dim):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """

    g_model = generator(input_z, out_channel_dim)
    d_model_real, d_logits_real = discriminator(input_real)
    d_model_fake, d_logits_fake = discriminator(g_model, reuse=True)
    
    #Label Smoothing, i.e. if you have two target labels: Real=1 and Fake=0, then for each incoming sample, if it is real, then replace the label with a random number between 0.7 and 1.2, and if it is a fake sample, replace it with 0.0 and 0.3 (for example).
    #Salimans et. al. 2016
    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, labels=tf.ones_like(d_model_real)*np.random.uniform(0.0, 0.9)))
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.zeros_like(d_model_fake)*np.random.uniform(0.0, 0.9)))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.ones_like(d_model_fake)))
    
    d_loss = d_loss_real + d_loss_fake
    
    
    return d_loss, g_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)

Tests Passed

Optimization¶

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    t_vars = tf.trainable_variables()
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    g_vars = [var for var in t_vars if var.name.startswith('generator')]
    
    #optimize
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)
        g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)
    
    return d_train_opt, g_train_opt


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)

Tests Passed

Neural Network Training¶

Show Output¶

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train¶

Implement train to build and train the GANs. Use the following functions you implemented:

model_inputs(image_width, image_height, image_channels, z_dim)
model_loss(input_real, input_z, out_channel_dim)
model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    input_real, input_z, learn_rate = model_inputs(data_shape[1], data_shape[2], data_shape[3], z_dim)
    d_loss, g_loss = model_loss(input_real, input_z, data_shape[3])
    d_opt, g_opt = model_opt(d_loss, g_loss, learning_rate, beta1)
    
    steps = 10
    
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                steps += 1
                
                #Normalize
                batch_images *= 2
                #Sample Noise for Generator
                batch_z = np.random.uniform(-1, 1, size=(batch_size, z_dim))
                #Optimizer 
                _ = sess.run(d_opt, feed_dict={input_real: batch_images, 
                                               input_z: batch_z, learn_rate: learning_rate})
                _ = sess.run(g_opt, feed_dict={input_real: batch_images, 
                                               input_z: batch_z, learn_rate: learning_rate})
                if steps % 10 == 0:
                    train_loss_d = d_loss.eval({input_z: batch_z, input_real: batch_images})
                    train_loss_g = g_loss.eval({input_z: batch_z})

                    print("Epoch {}/{}...".format(epoch_i + 1, epochs),
                          "Discriminator Loss: {:.4f}...".format(train_loss_d),
                          "Generator Loss: {:.4f}".format(train_loss_g))

                if steps % 100 == 0:
                    show_generator_output(sess, 16, input_z, data_shape[3], data_image_mode)

MNIST¶

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

batch_size = 16
z_dim = 100
learning_rate = 0.0004
beta1 = 0.1


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)

Epoch 1/2... Discriminator Loss: 1.3014... Generator Loss: 6.6167
Epoch 1/2... Discriminator Loss: 1.2600... Generator Loss: 1.6705
Epoch 1/2... Discriminator Loss: 0.7597... Generator Loss: 3.1203
Epoch 1/2... Discriminator Loss: 0.6636... Generator Loss: 3.8097
Epoch 1/2... Discriminator Loss: 0.8449... Generator Loss: 1.9559
Epoch 1/2... Discriminator Loss: 0.6363... Generator Loss: 4.4234
Epoch 1/2... Discriminator Loss: 0.6025... Generator Loss: 3.3128
Epoch 1/2... Discriminator Loss: 0.5546... Generator Loss: 3.6192
Epoch 1/2... Discriminator Loss: 0.6075... Generator Loss: 4.0272

Epoch 1/2... Discriminator Loss: 0.9289... Generator Loss: 5.9021
Epoch 1/2... Discriminator Loss: 0.5850... Generator Loss: 3.2355
Epoch 1/2... Discriminator Loss: 0.5624... Generator Loss: 3.4117
Epoch 1/2... Discriminator Loss: 0.6470... Generator Loss: 4.6036
Epoch 1/2... Discriminator Loss: 0.6388... Generator Loss: 3.1721
Epoch 1/2... Discriminator Loss: 0.5813... Generator Loss: 3.3534
Epoch 1/2... Discriminator Loss: 0.6802... Generator Loss: 4.0785
Epoch 1/2... Discriminator Loss: 0.5918... Generator Loss: 3.7308
Epoch 1/2... Discriminator Loss: 0.6539... Generator Loss: 3.9370
Epoch 1/2... Discriminator Loss: 0.7583... Generator Loss: 2.1247

Epoch 1/2... Discriminator Loss: 0.6639... Generator Loss: 3.7480
Epoch 1/2... Discriminator Loss: 0.7841... Generator Loss: 4.4122
Epoch 1/2... Discriminator Loss: 0.6369... Generator Loss: 3.2762
Epoch 1/2... Discriminator Loss: 0.6134... Generator Loss: 2.9869
Epoch 1/2... Discriminator Loss: 0.6039... Generator Loss: 3.7751
Epoch 1/2... Discriminator Loss: 0.6511... Generator Loss: 4.0164
Epoch 1/2... Discriminator Loss: 0.6740... Generator Loss: 4.0833
Epoch 1/2... Discriminator Loss: 0.5854... Generator Loss: 3.0759
Epoch 1/2... Discriminator Loss: 0.5580... Generator Loss: 3.2139
Epoch 1/2... Discriminator Loss: 0.6522... Generator Loss: 3.7598

Epoch 1/2... Discriminator Loss: 0.6933... Generator Loss: 3.9145
Epoch 1/2... Discriminator Loss: 0.6626... Generator Loss: 2.4683
Epoch 1/2... Discriminator Loss: 0.5774... Generator Loss: 3.3985
Epoch 1/2... Discriminator Loss: 0.5876... Generator Loss: 3.0516
Epoch 1/2... Discriminator Loss: 0.6314... Generator Loss: 2.2864
Epoch 1/2... Discriminator Loss: 0.5920... Generator Loss: 3.4765
Epoch 1/2... Discriminator Loss: 0.6374... Generator Loss: 2.4863
Epoch 1/2... Discriminator Loss: 0.5655... Generator Loss: 3.5066
Epoch 1/2... Discriminator Loss: 0.5796... Generator Loss: 2.6591
Epoch 1/2... Discriminator Loss: 0.6065... Generator Loss: 3.8335

Epoch 1/2... Discriminator Loss: 0.5944... Generator Loss: 2.9858
Epoch 1/2... Discriminator Loss: 0.5918... Generator Loss: 2.5011
Epoch 1/2... Discriminator Loss: 0.6167... Generator Loss: 3.0602
Epoch 1/2... Discriminator Loss: 0.6156... Generator Loss: 2.5156
Epoch 1/2... Discriminator Loss: 0.5752... Generator Loss: 2.5803
Epoch 1/2... Discriminator Loss: 0.5976... Generator Loss: 2.7336
Epoch 1/2... Discriminator Loss: 0.5806... Generator Loss: 2.7275
Epoch 1/2... Discriminator Loss: 0.6137... Generator Loss: 2.6083
Epoch 1/2... Discriminator Loss: 0.6198... Generator Loss: 2.4118
Epoch 1/2... Discriminator Loss: 0.6501... Generator Loss: 2.0055

Epoch 1/2... Discriminator Loss: 0.6619... Generator Loss: 2.2658
Epoch 1/2... Discriminator Loss: 0.6155... Generator Loss: 2.5966
Epoch 1/2... Discriminator Loss: 0.5780... Generator Loss: 3.2594
Epoch 1/2... Discriminator Loss: 0.5685... Generator Loss: 3.4639
Epoch 1/2... Discriminator Loss: 0.6708... Generator Loss: 2.6036
Epoch 1/2... Discriminator Loss: 0.5772... Generator Loss: 2.9101
Epoch 1/2... Discriminator Loss: 0.6715... Generator Loss: 2.2379
Epoch 1/2... Discriminator Loss: 0.5837... Generator Loss: 2.9408
Epoch 1/2... Discriminator Loss: 0.6064... Generator Loss: 2.7095
Epoch 1/2... Discriminator Loss: 0.6380... Generator Loss: 2.4066

Epoch 1/2... Discriminator Loss: 0.5749... Generator Loss: 3.2044
Epoch 1/2... Discriminator Loss: 0.6254... Generator Loss: 3.5452
Epoch 1/2... Discriminator Loss: 0.5892... Generator Loss: 3.2374
Epoch 1/2... Discriminator Loss: 0.7015... Generator Loss: 2.1361
Epoch 1/2... Discriminator Loss: 0.5657... Generator Loss: 2.9982
Epoch 1/2... Discriminator Loss: 0.5728... Generator Loss: 3.4896
Epoch 1/2... Discriminator Loss: 0.5751... Generator Loss: 2.8718
Epoch 1/2... Discriminator Loss: 0.5630... Generator Loss: 3.0678
Epoch 1/2... Discriminator Loss: 0.6084... Generator Loss: 2.5014
Epoch 1/2... Discriminator Loss: 0.5810... Generator Loss: 3.2987

Epoch 1/2... Discriminator Loss: 0.6058... Generator Loss: 3.4497
Epoch 1/2... Discriminator Loss: 0.6240... Generator Loss: 3.9604
Epoch 1/2... Discriminator Loss: 0.6334... Generator Loss: 2.3880
Epoch 1/2... Discriminator Loss: 0.6093... Generator Loss: 2.6209
Epoch 1/2... Discriminator Loss: 0.6393... Generator Loss: 3.6770
Epoch 1/2... Discriminator Loss: 0.5564... Generator Loss: 3.3268
Epoch 1/2... Discriminator Loss: 0.5744... Generator Loss: 3.2541
Epoch 1/2... Discriminator Loss: 0.8700... Generator Loss: 5.5112
Epoch 1/2... Discriminator Loss: 0.5759... Generator Loss: 2.9992
Epoch 1/2... Discriminator Loss: 0.6013... Generator Loss: 3.1171

Epoch 1/2... Discriminator Loss: 0.5846... Generator Loss: 2.7899
Epoch 1/2... Discriminator Loss: 0.5878... Generator Loss: 2.7861
Epoch 1/2... Discriminator Loss: 0.6221... Generator Loss: 2.2957
Epoch 1/2... Discriminator Loss: 0.6186... Generator Loss: 3.9788
Epoch 1/2... Discriminator Loss: 0.6354... Generator Loss: 2.2781
Epoch 1/2... Discriminator Loss: 0.6111... Generator Loss: 2.4637
Epoch 1/2... Discriminator Loss: 0.7180... Generator Loss: 1.9238
Epoch 1/2... Discriminator Loss: 0.5831... Generator Loss: 3.7088
Epoch 1/2... Discriminator Loss: 0.5527... Generator Loss: 2.9729
Epoch 1/2... Discriminator Loss: 0.6090... Generator Loss: 2.7290

Epoch 1/2... Discriminator Loss: 0.6448... Generator Loss: 2.5148
Epoch 1/2... Discriminator Loss: 0.6523... Generator Loss: 3.4896
Epoch 1/2... Discriminator Loss: 0.6361... Generator Loss: 2.1479
Epoch 1/2... Discriminator Loss: 0.6139... Generator Loss: 2.6206
Epoch 1/2... Discriminator Loss: 0.6162... Generator Loss: 2.4987
Epoch 1/2... Discriminator Loss: 0.5930... Generator Loss: 2.4167
Epoch 1/2... Discriminator Loss: 0.5541... Generator Loss: 3.3714
Epoch 1/2... Discriminator Loss: 0.8901... Generator Loss: 5.5349
Epoch 1/2... Discriminator Loss: 0.5699... Generator Loss: 2.8888
Epoch 1/2... Discriminator Loss: 0.5880... Generator Loss: 2.9115

CelebA¶

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

batch_size = 16
z_dim = 100
learning_rate = 0.0006
beta1 = 0.2


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 1

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)

Epoch 1/1... Discriminator Loss: 0.8959... Generator Loss: 3.6278
Epoch 1/1... Discriminator Loss: 0.5792... Generator Loss: 4.1814
Epoch 1/1... Discriminator Loss: 0.8200... Generator Loss: 1.7328
Epoch 1/1... Discriminator Loss: 0.5150... Generator Loss: 3.7084
Epoch 1/1... Discriminator Loss: 0.4916... Generator Loss: 3.9821
Epoch 1/1... Discriminator Loss: 0.4974... Generator Loss: 3.6824
Epoch 1/1... Discriminator Loss: 0.4861... Generator Loss: 4.5574
Epoch 1/1... Discriminator Loss: 0.4830... Generator Loss: 3.9070
Epoch 1/1... Discriminator Loss: 0.4735... Generator Loss: 4.2722

Epoch 1/1... Discriminator Loss: 0.4650... Generator Loss: 4.8683
Epoch 1/1... Discriminator Loss: 0.4771... Generator Loss: 3.9373
Epoch 1/1... Discriminator Loss: 0.4784... Generator Loss: 4.0915
Epoch 1/1... Discriminator Loss: 0.4582... Generator Loss: 4.4008
Epoch 1/1... Discriminator Loss: 0.5643... Generator Loss: 2.5403
Epoch 1/1... Discriminator Loss: 0.5724... Generator Loss: 2.4821
Epoch 1/1... Discriminator Loss: 0.5151... Generator Loss: 3.4557
Epoch 1/1... Discriminator Loss: 0.5088... Generator Loss: 2.9693
Epoch 1/1... Discriminator Loss: 0.5330... Generator Loss: 3.9567
Epoch 1/1... Discriminator Loss: 0.7139... Generator Loss: 5.2019

Epoch 1/1... Discriminator Loss: 0.6410... Generator Loss: 5.0973
Epoch 1/1... Discriminator Loss: 0.5155... Generator Loss: 3.6518
Epoch 1/1... Discriminator Loss: 0.5253... Generator Loss: 3.4016
Epoch 1/1... Discriminator Loss: 0.6671... Generator Loss: 2.6993
Epoch 1/1... Discriminator Loss: 0.5157... Generator Loss: 3.4921
Epoch 1/1... Discriminator Loss: 0.5147... Generator Loss: 3.6413
Epoch 1/1... Discriminator Loss: 0.4813... Generator Loss: 3.3208
Epoch 1/1... Discriminator Loss: 0.6237... Generator Loss: 4.4010
Epoch 1/1... Discriminator Loss: 0.4896... Generator Loss: 3.3151
Epoch 1/1... Discriminator Loss: 0.6945... Generator Loss: 5.2552

Epoch 1/1... Discriminator Loss: 0.5370... Generator Loss: 4.3267
Epoch 1/1... Discriminator Loss: 0.5328... Generator Loss: 2.7930
Epoch 1/1... Discriminator Loss: 0.5056... Generator Loss: 3.2624
Epoch 1/1... Discriminator Loss: 0.5117... Generator Loss: 3.5025
Epoch 1/1... Discriminator Loss: 0.5164... Generator Loss: 3.4097
Epoch 1/1... Discriminator Loss: 0.5226... Generator Loss: 3.2213
Epoch 1/1... Discriminator Loss: 0.5681... Generator Loss: 2.5757
Epoch 1/1... Discriminator Loss: 0.5222... Generator Loss: 2.9461
Epoch 1/1... Discriminator Loss: 0.5704... Generator Loss: 4.0829
Epoch 1/1... Discriminator Loss: 0.5325... Generator Loss: 3.3833

Epoch 1/1... Discriminator Loss: 0.5695... Generator Loss: 4.0921
Epoch 1/1... Discriminator Loss: 0.9873... Generator Loss: 6.1446
Epoch 1/1... Discriminator Loss: 0.5415... Generator Loss: 2.4762
Epoch 1/1... Discriminator Loss: 0.5459... Generator Loss: 3.5573
Epoch 1/1... Discriminator Loss: 0.5548... Generator Loss: 3.6768
Epoch 1/1... Discriminator Loss: 0.5395... Generator Loss: 3.7861
Epoch 1/1... Discriminator Loss: 0.6040... Generator Loss: 2.7150
Epoch 1/1... Discriminator Loss: 1.0158... Generator Loss: 5.9828
Epoch 1/1... Discriminator Loss: 0.5339... Generator Loss: 2.6733
Epoch 1/1... Discriminator Loss: 0.5283... Generator Loss: 2.7176

Epoch 1/1... Discriminator Loss: 0.5729... Generator Loss: 2.3153
Epoch 1/1... Discriminator Loss: 0.5448... Generator Loss: 2.7892
Epoch 1/1... Discriminator Loss: 0.7485... Generator Loss: 1.8531
Epoch 1/1... Discriminator Loss: 0.5335... Generator Loss: 2.9484
Epoch 1/1... Discriminator Loss: 0.8711... Generator Loss: 4.9164
Epoch 1/1... Discriminator Loss: 0.5207... Generator Loss: 2.7042
Epoch 1/1... Discriminator Loss: 0.5575... Generator Loss: 2.4942
Epoch 1/1... Discriminator Loss: 0.5308... Generator Loss: 3.0852
Epoch 1/1... Discriminator Loss: 0.5394... Generator Loss: 3.0268
Epoch 1/1... Discriminator Loss: 0.5496... Generator Loss: 2.4802

Epoch 1/1... Discriminator Loss: 0.5160... Generator Loss: 2.5462
Epoch 1/1... Discriminator Loss: 0.5385... Generator Loss: 3.5051
Epoch 1/1... Discriminator Loss: 0.5471... Generator Loss: 2.4159
Epoch 1/1... Discriminator Loss: 0.5620... Generator Loss: 2.5071
Epoch 1/1... Discriminator Loss: 0.5717... Generator Loss: 2.5130
Epoch 1/1... Discriminator Loss: 0.5389... Generator Loss: 3.0670
Epoch 1/1... Discriminator Loss: 0.5292... Generator Loss: 3.2755
Epoch 1/1... Discriminator Loss: 0.5680... Generator Loss: 2.7361
Epoch 1/1... Discriminator Loss: 0.5628... Generator Loss: 3.8917
Epoch 1/1... Discriminator Loss: 0.5375... Generator Loss: 2.7423

Epoch 1/1... Discriminator Loss: 0.5648... Generator Loss: 2.8092
Epoch 1/1... Discriminator Loss: 0.5208... Generator Loss: 2.8177
Epoch 1/1... Discriminator Loss: 0.5333... Generator Loss: 3.1625
Epoch 1/1... Discriminator Loss: 0.5385... Generator Loss: 3.9560
Epoch 1/1... Discriminator Loss: 0.5288... Generator Loss: 2.9408
Epoch 1/1... Discriminator Loss: 0.5614... Generator Loss: 2.4451
Epoch 1/1... Discriminator Loss: 0.5501... Generator Loss: 3.6566
Epoch 1/1... Discriminator Loss: 0.5398... Generator Loss: 3.4950
Epoch 1/1... Discriminator Loss: 0.5303... Generator Loss: 2.9648
Epoch 1/1... Discriminator Loss: 0.5680... Generator Loss: 2.4651

Epoch 1/1... Discriminator Loss: 0.5559... Generator Loss: 2.3827
Epoch 1/1... Discriminator Loss: 0.5091... Generator Loss: 3.0505
Epoch 1/1... Discriminator Loss: 0.5410... Generator Loss: 2.9456
Epoch 1/1... Discriminator Loss: 0.5781... Generator Loss: 2.3701
Epoch 1/1... Discriminator Loss: 0.5145... Generator Loss: 2.7550
Epoch 1/1... Discriminator Loss: 0.5328... Generator Loss: 2.2721
Epoch 1/1... Discriminator Loss: 0.5700... Generator Loss: 2.4613
Epoch 1/1... Discriminator Loss: 0.5674... Generator Loss: 2.5288
Epoch 1/1... Discriminator Loss: 0.5452... Generator Loss: 2.6119
Epoch 1/1... Discriminator Loss: 0.6297... Generator Loss: 4.0412

Epoch 1/1... Discriminator Loss: 0.5312... Generator Loss: 2.6895
Epoch 1/1... Discriminator Loss: 0.5148... Generator Loss: 2.8668
Epoch 1/1... Discriminator Loss: 0.6318... Generator Loss: 2.3776
Epoch 1/1... Discriminator Loss: 0.5376... Generator Loss: 2.6848
Epoch 1/1... Discriminator Loss: 0.5575... Generator Loss: 2.1999
Epoch 1/1... Discriminator Loss: 0.5590... Generator Loss: 2.5374
Epoch 1/1... Discriminator Loss: 0.5569... Generator Loss: 2.9384
Epoch 1/1... Discriminator Loss: 0.5417... Generator Loss: 2.5073
Epoch 1/1... Discriminator Loss: 0.5325... Generator Loss: 2.9616
Epoch 1/1... Discriminator Loss: 0.5275... Generator Loss: 2.8082

Submitting This Project¶

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.